Abstract: The paper examines whether a long-run relationship between fifinancial development and economic growth exists employing panel integration and cointegration techniques for a dynamic heterogeneous panel of 15 OECD and 50 nonOECD countries over the period 1975–2000. Three different measures of fifinancial deepening are used to capture the variety of different channels through which fifinancial development can affect growth. Our fifindings support the existence of a single long-run equilibrium relation between fifinancial deepening, economic growth and a set of control variables. Further, the evidence points to a bi-directional causality between fifinancial deepening and growth. JEL no. O11, O16, C33 Keywords: Financial development; growth; panel cointegration; panel causality
Academic research on the fifinance-growth nexus dates back at least to Schumpeter (1911) who emphasized the positive role of fifinancial development on economic growth.1 The contribution of the fifinancial markets to growth has received considerable attention with the emergence of the endogenous growth theory,2 however, the related literature started expanding vigorously especially after the seminal study of King and Levine (1993a,1993b),3 which revived the interest and gave a boost to a plethora of research.
ther, to examine the direction of causality between fifinancial development and economic growth.
When it comes to time series studies, although one is able to identify the direction of causality and the nature of the I(1) variables is taken into account in the estimation techniques, there is always the possibility of misleading standard tests and unreliable results due to short length of data sets (Pierse and Shell 1995).
Finally, panel data studies make a serious attempt to control for possible shortcomings of the previous two methodologies by accounting for other determinants of growth to avoid potential biases induced by omitted variables, correcting for simultaneity using instrumental variables and GMM dynamic panel estimator, and controlling for unobserved country-specifific effects. However, they ignore the integration properties of their data. Therefore, it is not clear whether they eventually estimate a long-run equilibrium relationship between fifinance and growth or a spurious one offering thus misleading conclusions.
Consequently, the purpose of the present paper is to re-examine the nature of the fifinance-growth relationship and provide better empirical insights by considering the application of novel econometric techniques— panel cointegration and panel estimation—that exploit the full information of the data while correcting at the same time many of the shortcomings mentioned above.
To our knowledge, there is only one study so far in the fifinance-growth nexus literature that employs panel data cointegration techniques to investigate the nature of this relationship. Christopoulos and Tsionas (2004), use panel cointegration analysis to examine whether a long-run relationship between fifinancial development and economic growth exists for 10 developing countries over the period 1970–2000. Their fifindings are supportive to a unique cointegrating vector between growth, fifinancial development,investment share, and inflflation, and to unidirectional causality from fifinancial depth to growth. However, this study limits its attention only to few developing countries and employs only one measure of fifinancial deepening.
The present paper aims to contribute to the relevant literature in the following ways:
i) On the econometric front, we make use of panel unit root (Im et al.2003) and panel cointegration tests (Pedroni 1999) allowing for heterogeneity in coeffificients and dynamics across units, which enable us to determine the long-run structure of the fifinance-growth relationship avoiding wellknown problems that occur in using traditional (time series) cointegration testing (low-power due to small samples). The cointegrating vectors are estimated using dynamic OLS (DOLS) procedure, which allows for consistent and effificient estimators of the long-run relationship, deals with the endogeneity of the regressors and takes into account the integration and cointegration properties of the data.
ii) We attempt to resolve the issue of causality, i.e., whether better functioning fifinancial markets exert a causal inflfluence on growth or vice versa following the methodology of Pesaran et al. (1999).
iii) We employ various measures of fifinancial development in order to quantify the impact of fifinancial depth on growth and, further, to examine the sensitivity of the results. The majority of the previous studies use a single measure of fifinancial development and base their fifindings solely on it. But no measure is perfect, and further, there are many channels through which fifinancial deepening could impact on growth and these channels cannot be explored by employing only a single indicator.
iv) To control for possible omitted variable bias, we include a complete conditioning information set consisting of variables borrowed from the relevant literature.
v) We use a large and heterogeneous sample of 65 countries (15 OECD and 50 non-OECD) over the period 1975–2000.
Our fifindings indicate that there is clearly a positive association between fifinancial deepening and economic growth. Further, the results support a bidirectional causality between fifinancial deepening and growth for the panels of OECD and non-OECD countries. We conclude that policies aiming at improving fifinancial markets (economic growth) will have, in the long-run,a signifificant effect on economic growth (fifinancial development).
The remainder of the paper is organized as follows: Section 2 presents the model under estimation and data. The econometric methodology is introduced in Section 3. Section 4 estimates the econometric specifification and discusses the empirical results. Section 5 concludes.
EasyDue™ 支持PayPal, AliPay, WechatPay, Taobao等各种付款方式!
E-mail: email@example.com 微信:easydue